基于特征的机载导弹阵地抢修性聚类评估方法  

Clustering evaluation method based on characteristic for the combat resilience of airborne missile camp

作　　者：陈希林[1] 肖明清[1] 王学奇[1] 

机构地区：[1]空军工程大学工程学院

出　　处：《系统工程与电子技术》2008年第6期1093-1097,共5页Systems Engineering and Electronics

基　　金：国防“十一五”重点预研项目(51317030103);空军工程大学优秀博士论文创新基金(BC06005)资助课题

摘　　要：结合机载导弹阵地战场抢修决策平台的开发,建立了机载导弹阵地装备抢修性评价体系,同时为了合理引入抢修性评价过程中的专家分歧,提出了基于特征的聚类评估方法。该方法首先将评估矩阵进行奇异值分解,提取对矩阵特征贡献率大的奇异值进行分解逆运算,然后选取欧氏距离和离差平方和作为层次式聚类方法的分类指标,对数据进行聚类分析。文末应用上述评价体系和评估方法对某型机载导弹阵地典型装备进行了抢修性评估。应用表明:适当引入专家分歧使得聚类分析结果趋向合理化。所提评估方法对群决策问题适用。Aiming at the development of decision support platform for BDAR in airborne missile camp （DSP-BDAR）, the paper sets up an evaluation system of combat resilience of airborne missile camp materiel, and puts forward a clustering evaluation method based on characteristics in order to depress the expert diver-gence during the evaluation of combat resilience. In the evaluation method, singular value decomposition is put on the evaluation matrix first, and the singulars which have more contribution to matrix characteristics are cho-sen to attend the converse decomposing operation of the evaluation matrix. And then the Euclidean metric and inner squared distance are chosen as the clustering indexes to carry out a clustered analysis. In the end, the paper gives out an evaluation example on typical equipments in the airborne missile camp of a certain aircraft by using the proposed evaluation system and evaluation method. The demonstration shows that it is the introduction of the expert divergence properly that contributes to the rationalization of the evaluation result. The method is suitable for the group deeision-making problem.

关 键 词：战场抢修 抢修性 奇异值分解 层次式聚类方法 导弹装备 

分 类 号：TP39[自动化与计算机技术—计算机应用技术] E82[自动化与计算机技术—计算机科学与技术]